When your website is down for an hour, the amount you lose is often greater than your entire annual hosting budget. Yet most businesses have never calculated this figure — and because they haven't, they cannot know how much their infrastructure actually deserves to be invested in.
Short answer: how is downtime cost calculated?
Downtime cost is the sum of lost hourly revenue and idle staff cost, multiplied by the duration of the outage. The base formula is: (annual revenue ÷ annual working hours) + (hourly cost of affected staff). This is the baseline figure, before indirect items such as reputational damage and customer acquisition cost are added.
The four cost components of downtime
Treating downtime purely as "sales I couldn't make during that hour" captures only a small part of the picture. The real cost consists of four components:
1. Direct revenue loss
For e-commerce this is the most visible item: orders that cannot be completed while you are offline. But service businesses are not exempt either. A corporate site whose contact form is broken loses every enquiry in that window — and most of those enquiries never return, because the visitor simply moves to the next search result.
2. Idle staff cost
Every employee who cannot do their job while the system is down is a cost that keeps accruing. Sales staff unable to enter orders, operations locked out of the panel, support unable to answer customers. This item is routinely forgotten in calculations, yet it is often comparable in size to the direct revenue loss.
3. Recovery and response cost
The time your technical team (or emergency external support) spends resolving the issue, restoration work if data was lost, and the verification testing that follows. Emergency intervention is always more expensive than planned maintenance.
4. Reputation and trust erosion
This is the hardest to measure and the longest lasting. A first-time visitor who lands on an error page forms a judgement about your brand within seconds and usually does not try again. Among existing customers, repeated outages directly influence renewal decisions.
What uptime percentages mean in real hours
Hosting and server providers advertise figures such as "99.9% uptime". The gaps between these percentages look small but produce very different outcomes in practice:
| Uptime commitment | Allowed downtime per year | Per month | What it means in practice |
|---|---|---|---|
| 99% | 3 days 15 hours | ~7 hours | Inadequate for serious workloads |
| 99.5% | 1 day 19 hours | ~3.6 hours | Borderline for a corporate site |
| 99.9% | 8 hours 45 minutes | ~43 minutes | Common standard for business services |
| 99.95% | 4 hours 22 minutes | ~22 minutes | Recommended level for e-commerce |
| 99.99% | 52 minutes | ~4.4 minutes | Outage-sensitive critical systems |
The difference between 99% and 99.9% looks like a single decimal place; in reality it is the difference between 3.5 days and 9 hours of annual downtime. Above a certain revenue level, the monetary value of that gap comfortably exceeds the price difference between the two service tiers.
What to check in an SLA
An uptime commitment only means something if the contract behind it is specific. Ask for written answers to these questions:
- How is uptime measured? Is the window monthly or annual, and from which measurement point?
- Is planned maintenance included? In most contracts it is excluded — the maximum maintenance window should be stated explicitly.
- What happens if the commitment is missed? Is there a credit or refund mechanism, how is it claimed, and is it applied automatically?
- Is there a response time (RTO) commitment? "Keeping it running" and "how fast it is restored after a failure" are two different promises.
- How often are backups taken, and what is the restore time? What matters is not that a backup exists, but how long restoration takes.
Calculate your own downtime cost
- Find your hourly revenue. Divide annual revenue by actual annual working hours (e.g. 45 hours a week, roughly 2,300 hours a year).
- Apply your digital channel share. What percentage of revenue arrives through the website, forms, panel or online store? Multiply hourly revenue by that share.
- Add idle staff cost. Number of affected employees × average fully loaded hourly cost.
- Apply a peak-hour multiplier. An outage during business hours or a campaign period can cost two to three times the average.
- Convert to annual risk. Multiply your hourly cost by the annual downtime allowed at your current uptime level. That figure is the benchmark for your infrastructure investment.
Netişlem expert view: the three most common causes we see in the field
Across the on-site and remote interventions we carry out, the overwhelming majority of outages come not from exotic technical faults but from three ordinary causes:
First, an expired domain name or SSL certificate. The server is technically up, but the site is unreachable — and because nobody is tracking it, discovery takes hours.
Second, resource exhaustion. The business grows, traffic and data grow, but the plan stays the same for years. The outage does not arrive suddenly; it announces itself first as slowness — an early warning that is usually ignored.
Third, uncontrolled updates. Plugin, theme or version updates applied directly to production without a staging environment. On a system whose backups have never been verified, this turns an outage into data loss.
All three share one trait: none of them requires an expensive solution. They require regular monitoring.
Frequently asked questions
Is 99.9% uptime enough?
It is sufficient for corporate brochure sites and most B2B service sites; it corresponds to roughly 9 hours of downtime per year. E-commerce sites taking continuous orders, booking systems and customer portals should target 99.95% or above.
How do I include reputational damage in the calculation?
Its direct monetary value is hard to measure. A practical approach is to multiply the number of lost prospects by your customer acquisition cost, which gives you a lower bound on the real impact.
Does planned maintenance count as downtime?
It depends on the contract. In most SLAs, maintenance announced in advance and performed within a defined window is excluded from the uptime calculation. This clause should be explicit, and the maximum maintenance window should be stated.
I have backups — am I still at risk?
Yes. The existence of a backup is not enough on its own. What matters is the restore time, and whether restoration is regularly tested. A backup that has never been tested is an assumption, not a guarantee.
How should I monitor my uptime?
Using an independent monitoring service is the healthiest approach; measuring from outside, in addition to the provider's own reporting, also matters for contract enforcement. Monitoring should cover not just the homepage but critical flows such as forms and the checkout step.
Conclusion
Calculating downtime cost moves infrastructure decisions from "how much does this cost" to "how much am I putting at risk". Those two questions produce very different budgets — and the second one is the correct question.
If you would like us to assess the real risk level of your current infrastructure, get in touch to request an on-site or remote evaluation. You can also review our SLA commitment and our corporate hosting solutions.